Model-Driven Development of Complex Software: a Research Roadmap Robert France, Bernhard Rumpe

Model-Driven Development of Complex Software: a Research Roadmap Robert France, Bernhard Rumpe

Model-driven Development of Complex Software: A Research Roadmap Robert France, Bernhard Rumpe Robert France is a Professor in the Department of Computer Science at Colorado State University. His research focuses on the problems associated with the development of complex software systems. He is involved in research on rigorous software modeling, on providing rigorous support for using design patterns, and on separating concerns using aspect-oriented modeling techniques. He was involved in the Revision Task Forces for UML 1.3 and UML 1.4. He is currently a Co-Editor-In-Chief for the Springer international journal on Software and System Modeling, a Software Area Editor for IEEE Computer and an Associate Editor for the Journal on Software Testing, Verification and Reliability. Bernhard Rumpe is chair of the Institute for Software Systems Engineering at the Braunschweig University of Technology, Germany. His main interests are software development methods and techniques that benefit form both rigorous and practical approaches. This includes the impact of new technologies such as model-engineering based on UML-like notations and domain specific languages and evolutionary, test-based methods, software architecture as well as the methodical and technical implications of their use in industry. He has furthermore contributed to the communities of formal methods and UML. He is author and editor of eight books and Co-Editor-in-Chief of the Springer International Journal on Software and Systems Modeling (www.sosym.org). Future of Software Engineering(FOSE'07) 0-7695-2829-5/07 $20.00 © 2007 Model-driven Development of Complex Software: A Research Roadmap Robert France Bernhard Rumpe Department of Computer Science Software Systems Engineering Institute Colorado State University Faculty for Mathematics and Computer Science Fort Collins, CO 80523 Braunschweig University of Technology [email protected] Braunschweig, Germany http://www.sse.cs.tu-bs.de Abstract Bridging the gap using approaches that require extensive handcrafting of implementations gives rise to accidental The term Model-Driven Engineering (MDE) is typically complexities that make the development of complex soft- used to describe software development approaches in which ware difficult and costly. To an extent, handcrafting com- abstract models of software systems are created and system- plex software systems can be likened to building pyramids atically transformed to concrete implementations. In this in ancient Egypt. We marvel at these software implemen- paper we give an overview of current research in MDE and tations in much the same way that archaeologists marvel at discuss some of the major challenges that must be tackled the pyramids: The wonder is mostly based on an apprecia- in order to realize the MDE vision of software development. tion of the effort required to tackle the significant accidental We argue that full realizations of the MDE vision may not complexities arising from the use of inadequate technolo- be possible in the near to medium-term primarily because gies. of the wicked problems involved. On the other hand, at- The growing complexity of software is the motivation tempting to realize the vision will provide insights that can behind work on industrializing software development. In be used to significantly reduce the gap between evolving particular, current research in the area of model driven software complexity and the technologies used to manage engineering (MDE) is primarily concerned with reducing complexity. the gap between problem and software implementation do- mains through the use of technologies that support system- atic transformation of problem-level abstractions to soft- 1. Introduction ware implementations. The complexity of bridging the gap is tackled through the use of models that describe com- Advances in hardware and network technologies have plex systems at multiple levels of abstraction and from a paved the way for the development of increasingly per- variety of perspectives, and through automated support for vasive software-based systems of systems that collaborate transforming and analyzing models. In the MDE vision of to provide essential services to society. Software in these software development, models are the primary artifacts of systems is often required to (1) operate in distributed and development and developers rely on computer-based tech- embedded computing environments consisting of diverse nologies to transform models to running systems. devices (personal computers, specialized sensors and ac- Current work on MDE technologies tends to focus on tuators), (2) communicate using a variety of interaction producing implementation and deployment artifacts from paradigms (e.g., SOAP messaging, media streaming), (3) detailed design models. These technologies use models to dynamically adapt to changes in operating environments, generate significant parts of (1) programs written in lan- and (4) behave in a dependable manner [26, 62]. De- guages such as Java, C++, and C] (e.g., see Compuware’s spite significant advances in programming languages and OptimalJ, IBM’s Rational XDE package, and Microsoft’s supporting integrated development environments (IDEs), Visual Studio), and (2) integration and deployment artifacts developing these complex software systems using current such as XML-based configuration files and data bridges code-centric technologies requires herculean effort. used for integrating disparate systems (e.g., see [25]). A significant factor behind the difficulty of develop- Attempts at building complex software systems that dy- ing complex software is the wide conceptual gap between namically adapt to changes in their operating environments the problem and the implementation domains of discourse. has led some researchers to consider the use of models dur- Future of Software Engineering(FOSE'07) 0-7695-2829-5/07 $20.00 © 2007 ing runtime to monitor and manage the executing software. challenges in the areas of modeling languages, separation Early work in this emerging MDE area was presented at a of concerns, and model manipulation and management, re- MODELS 2006 Workshop on runtime models [8]. spectively. Section 7 also includes a discussion on oppor- We envisage that MDE research on runtime models will tunities for using models during runtime. We conclude in pave the way for the development of environments in which Section 8 by outlining an idealistic vision of an MDE envi- change agents (e.g., software maintainers, software-based ronment. agents) use runtime models to modify executing software in a controlled manner. The models act as interfaces that 2. The Value of Modeling change agents can use to adapt, repair, extend, or retrofit software during its execution. In our broad vision of MDE, In this paper, a model is an abstraction of some aspect of models are not only the primary artifacts of development, a system. The system described by a model may or may not they are also the primary means by which developers and exist at the time the model is created. Models are created to other systems understand, interact with, configure and mod- serve particular purposes, for example, to present a human ify the runtime behavior of software. understandable description of some aspect of a system or A major goal of MDE research is to produce technolo- to present information in a form that can be mechanically gies that shield software developers from the complexities analyzed (e.g., see [54, 30]). of the underlying implementation platform. An implemen- It may seem that work on MDE centers on the devel- tation platform may consist of networks of computers, mid- opment and use of the popular modeling language, UML dleware, and libraries of utility functions (e.g., libraries of (the Unified Modeling Language) [59]. The UML standard- persistence, graphical user interface, and mathematical rou- ization effort has played a vital role in bringing together a tines). In the case of MDE research on runtime models, the community that focuses on the problem of raising the level goal is to produce technologies that hide the complexities of of abstraction at which software is developed, but research runtime phenomena from agents responsible for managing around other modeling languages is contributing valuable the runtime environment, and for adapting and evolving the MDE concepts, techniques, tools and experience. In this software during runtime. paper, MDE encompasses all research pertaining to the use Realizing the MDE vision requires tackling a wide range of software models. of very difficult interrelated social and technical problems Non-UML modeling approaches that are included in our that has been the focus of software engineering research use of the MDE term include specifying systems using over the last three decades. For this reason, we consider formal specification languages such as Alloy [32] and B the problem of developing MDE technologies that automate [2], modeling and analyzing control system software us- significant portions of the software lifecycle to be a wicked ing the math-based, high-level programming language Mat- problem. A wicked problem has multiple dimensions that lab/Simulink/Stateflow (e.g., see [34]), analyzing perfor- are related in complex ways and thus cannot be solved by mance, load, safety, liveness, reliability, and other system cobbling solutions to the different problem dimensions (see properties using specialized modeling techniques (e.g., see definition of “wicked problem” in Wikipedia). Solutions to [40]), and

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    18 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us